This project provides a template implementation of a Model Context Protocol (MCP) server, integrating with Mem0 to offer AI agents persistent, semantically indexed long-term memory. It's designed for developers building custom MCP servers or as a reference for AI coding assistants, enabling agents to store, retrieve, and search memories efficiently.

How It Works

The server implements three core memory management functions: save_memory for storing and indexing, get_all_memories for retrieving all data, and search_memories for semantic retrieval. It leverages Mem0 for vector storage and semantic indexing, adhering to Anthropic's best practices for MCP server development, ensuring compatibility with MCP-compatible clients.

Quick Start & Requirements

• Install: Clone the repository and install dependencies using uv pip install -e ..
• Prerequisites: Python 3.12+, PostgreSQL (Supabase recommended) for vector storage, LLM API keys (OpenAI, OpenRouter, Ollama), and Docker (recommended).
• Setup: Create a .env file from .env.example and configure environment variables.
• Running: Use uv run src/main.py for SSE transport or configure clients for stdio transport. Docker images can be built and run similarly.
• Docs: Configuration details and integration examples for various clients (Claude Desktop, Windsurf, n8n) are provided.

Highlighted Details

• Supports both SSE (Server-Sent Events) and stdio transport protocols.
• Configurable LLM providers (OpenAI, OpenRouter, Ollama) and embedding models.
• PostgreSQL/Supabase integration for persistent vector storage.
• Serves as a foundational template for building more complex MCP servers with custom tools and resources.

Maintenance & Community

The repository is maintained by coleam00. No specific community channels or roadmap links are provided in the README.

Licensing & Compatibility

The repository does not explicitly state a license in the README. Compatibility for commercial use or closed-source linking is not specified.

Limitations & Caveats

The project is presented as a template and reference implementation. Specific performance benchmarks, detailed error handling, or production-readiness assessments are not included. The absence of an explicit license may pose a barrier to commercial adoption.